MicroZig bans LLM use for code, prose, editing, translation and bug finding

The Zig Embedded Group's code of conduct now carries a "Strict No LLM / No AI Policy" covering its GitHub organization and Discord server. It rules out LLM-generated code or prose, paraphrasing LLM output, LLMs for editing "including fixing spelling or grammatical errors," translation, and brainstorming whose results are then shared. It also bans "LLMs for finding bugs" and "talking about use of chatbot/LLM services." A comment on an issue written with a chatbot's advice is "unwelcome."

On the Hacker News thread (55 points, 73 comments about 22 hours in), Matthew Knight, who describes himself as the most active maintainer, said the choice was his: "one approach cannot coexist with the other," and he does not mind if someone makes an LLM fork. Commenters split on the bug-finding clause. Some read it literally, others as aimed at unreviewed machine-generated reports. The policy text does not say which, and the part of the thread we read has no maintainer clarification. One commenter, katbyte, said they forked and fixed a bug rather than open a pull request after seeing the rule; that is a single user's account.

Why it matters: QGIS banned AI contributions earlier this month on review-capacity grounds, and this policy reaches further: it covers issues, discussion and translation, not only code. If an agent of yours files bugs, drafts pull request text or answers in project chats upstream, check each project's policy first. Under this one, an agent-assisted bug report is a violation even if a person reviewed it.